"Unless we find a way to bind these awful Forms to our
Embrace we shall perish."

- Jerusalem 82.3-4 (E 239)[1]

Few artists have linked their arguments to their methods of production with the same conscious intent as the engraver, poet, and painter William Blake. In his most ambitious declarations, Blake hoped to foster nothing less than a full-scale media revolution that would promote a more direct, personal relationship between artists and audiences. Best known today for "illuminated printing," a form of relief etching he created in the late 1780s (E 693), Blake often worked proudly and determinedly as his own bookseller, publisher, and editor, shaping his works as he saw fit. In An Island in the Moon (c. 1783-85), Blake offered his first vision of a new method of printing that would transform his world: "I would have all the writing Engraved instead of Printed & at every other leaf a high finishd print all in three Volumes folio, & sell them a hundred pounds a piece. they would Print off two thousand." The radical political implications of this media revolution are seen when a character associated with Blake's wife Catherine responds with the air of a sans-culotte: "whoever will not have them will be ignorant fools & will not deserve to live" (E 465).

Even after he perfected illuminated printing, Blake continued to experiment with a vast range of different media: illustrated and illuminated manuscripts, intaglio etchings, emblem books, the marginal designs he produced for Young and Gray, the white-line etching for Blair, his unique fresco, and the original fusion of word and image in the illustrations to the Book of Job and many other works. But if Blake desired to reform society around these new forms of media, it was a future that never came to pass. He discovered time and again how dependent he was upon existing networks of media production and distribution. It is likely that at the same moment when the 1793 Prospectus blamed publishers for the "poverty and obscurity" of "the Artist, the Poet, [and] the Musician" who lacked the "means to propagate" their own works (E 693), Blake's books were on display at the shop of the radical bookseller Joseph Johnson (Davies 216-17). Ironically, his most productive period in illuminated printing coincides with his most successful period as a commercial engraver. As much as Blake thought about different forms of media production, in the end his images of editing and publishing were often vague. It is unclear in the Island fantasy who the "they" were who would print and sell these engraved books and which "them" Catherine would execute for lack of taste. In Marriage's "Printing house of Hell," Blake's "Unnam'd forms" are passively "reciev'd by Men" who put them into "the forms of books" (E 40). As seen in Poetical Sketches (1783), possibly Tiriel (c. 1789), The French Revolution (1791), and his plan to print some version of The Four Zoas (1797-1808?) to great profits (E 726), Blake continually flirted with the possibilities of publishing in more conventional forms of print. Blake also published his visual art in exhibitions of all types. In addition to the Royal Academy, Blake's illustrations to Young and Blair were displayed at bookseller shops alongside other illustrated book projects, and he took the more radical step of holding his own exhibition in 1809-10 only when he was denied the use of more traditional venues, printing the Descriptive Catalogue and the Prospectus for The Canterbury Pilgrims in letterpress without ambivalence. Blake's failure to publish also led to some of his most intriguing work. The manuscripts Island, Tiriel, The Four Zoas, and the Genesis illuminations are complex works that exploit their material form in profound ways, while the notebooks and even the marginalia consist of important visual and verbal play.

I

Blake's extensive use of different media forms has meant that the first task of every editor has been to remediate this work—to translate it to a new medium, which irrevocably changes its form, context, circulation, and meaning. It is for this reason that professional Blake editing began with John Sampson's recognition of the "disservice" (v) an editor must do to Blake's intentions. Invariably, no edition can convey all that Blake intended. While it may sometimes seem that the selection of a medium would be only one among many editorial choices, remediation is a kind of superstructure that conditions all subsequent editorial decisions. The concept of remediation was formulated by Jay David Bolter and Richard Grusin, who adapted it from the media theory of Marshall McLuhan.[2] While the content metaphor of remediation has been criticized as positivistic,[3] its relevancy to editing print and digital media has recently been explored by Diana Kichuk, Kathryn Sutherland, and Alan Liu. In each of these studies, theories of remediation serve to remind writers, readers, and editors of their place in what Liu calls an "encounter zone" of shifting media protocols (Liu par. 1). For editors and readers of Blake, this scene resembles nothing so much as the title page of The [First] Book of Urizen, where Urizen sits reading/writing/drawing/transcribing the different media codes before him.[4] As this title page makes clear, Urizen's Newtonian search for "a solid without fluctuation" (E 71) is also a search for pure linguistic codes that dwell apart from bibliographic codes, the seamlessly remediated stream of data. Kichuk has identified four paradoxical claims of remediation apposite to Urizen's strivings:

(1) The remediation is the "real thing" or a clone with a primary focus on the old medium.
(2) The remediation seeks to improve the old medium.
(3) The old medium is intentionally fashioned or changed.
(4) The old medium is "absorbed" into the new media without a trace. (292)

The contradictions of remediations are manifest in all editions of Blake's works, be they letterpress editions, print facsimiles, book reproductions, catalogues, digital archives, and, arguably, even exhibitions. Editions claim to represent the reality of Blake's work, even while improving it, and as the work becomes a new thing, it remains itself all the while. Such paradoxes are familiar to editorial theory, but the preponderance of remediation in our culture threatens to veil its contradictions. Today, readers of Blake will encounter the poet in a range of different media. Most readers of this electronic collection will use their computers to consult the William Blake Archive while cross-referencing their own dog-eared copies of Erdman, Bentley, or Keynes, the Princeton-Blake Trust facsimiles of the illuminated books, Blake Books, Blake Records, and other print materials, many of which also exist in digital forms. This host of different forms allows readers to see Blake in numerous ways, fulfilling the injunction of the Santa Cruz Blake Study Group "to triangulate Blake's text through as many editions and editorial theories as one can lay hands on" (15).

Proposed by Justin Van Kleeck, Editing Blake hopes to aid in this readerly triangulation by bringing together the work of both established and younger Blake editors and scholars reflecting on the problems and possibilities of editing Blake in the current media environment. W. H. Stevenson, Mary Lynn Johnson, and David Fuller have all edited successful print editions: The Poems of William Blake (Longman: 1971, 1989, 2007), Blake's Poetry and Designs (with John Grant, Norton: 1979, 2008), and William Blake: Selected Poetry (Longman: 2000, 2008). These editions have not only thrived in the shadow of Keynes, Erdman, and Bentley, but given the fact that all three editions have been recently revised, they remain vibrant print editions of the digital age. The essays of these editors explicate their editorial methodologies and their principles of revision, as well as detailing the ways in which material, social, and economic realities have impacted their volumes. Similarly, the essays suggest that editors of codex editions of Blake today can ignore neither the editorial implications of digital media nor the wealth of resources that digital media makes available to readers of all levels. The second set of contributors, Justin Van Kleeck, Rachel Lee, J. Alexandra McGhee, and Wayne C. Ripley, have all worked as project assistants to the Blake Archive and received their graduate training from its editors. Their essays are in more direct dialogue with the Archive as it expands its editorial labors to Blake's works beyond the illuminated books to include the manuscripts and illustrations of other authors, investigating how these new forms have or have not required the Archive to revise its editorial theory and practice. Lee and McGhee provide a useful and detailed history of the Archive up through its present work in preparing Blake's manuscript satire, An Island in the Moon. Drawing on his dissertation, which is a thoughtful and comprehensive survey of the problems of editing Blake,[5] Van Kleeck details the editorial history of Blake's heavily revised illuminated manuscript The Four Zoas, including the electronic edition he prepared for his dissertation that will become the basis for the edition at the Archive. Ripley uses the theories of D. F. McKenzie and Jerome McGann to pursue a Blakean notion of social-text editing for Blake's illustrations of other authors. Together, these six essays show how past and present experiences of editing Blake contain much that is instructive and fully applicable to editing in today's media environment.

In its focus on editorial practice and achievement, Editing Blake can be seen as a companion of the 2005 Praxis volume, Digital Designs on Blake. Edited by Ron Broglio, this volume explored "how new media representation of William Blake's work provides a heuristic for another mode of inquiry into Blake's complex verbal and visual texts" ("About This Volume"). Instead of employing the model of the archive common to electronic editions like the Blake Archive, Broglio offers "immersive textuality" as a model for engaging literary works from within ("Living" par. 5). Many of these forms of representation involve the creative exploitation of ideas implicit in Blake's own text or philosophy and use tools, such as MOO, Flash animation, and Ivanhoe, to draw them out. Such approaches pursue Jerome McGann's idea that all editorial ventures are species of interpretative "deformations" (Radiant 116). The deforming role of the editor is a point acknowledged by all the contributors to Editing Blake, but Morris Eaves's warning against the ever-present threat of "change and obsolescence" in digital media is worth noting ("Crafting" par. 29). Most of the electronic heuristics explored by Digital Designs have had a very brief shelf life and limited popular appeal as a means of presenting and exploring Blake's work, and it remains the case that print media and the self-declared "conservative" Blake Archive are the primary means by which almost all of Blake's readers engage his works ("Crafting" par. 25).

To say this is not to criticize the innovative work of Digital Designs, which makes no claims about editing Blake, but to recognize the still unwritten future of scholarly editing and the uncertain demands of future readers in our complex media environment. It is highly probable that Digital Designs will offer much in terms of editing Blake in an interactive Web 2.0 environment, but potential interactive models of editing still must reconcile themselves to questions of textual and pictorial accuracy. Peter Robinson has faulted scholarly editions for not moving "beyond the model of print technology," arguing that "dynamic interactivity will change the relationship of the reader to the text" ("Where We Are" pars. 1 and 9). However, Robinson's notion of dynamic interactivity is not exemplified by MOO spaces but more sophisticated means of presenting and analyzing traditional editorial data, such as the collation of variants and manuscript groupings that would be of little use to the general reader. Which type of interactivity, if either, will represent the editorial future? Robinson, who clearly believes in the superiority of digital scholarly editions over print ("Current Issues" par. 5), has recognized how the difficulty of producing digital editions has contributed to the resistance of scholars, publishers, and readers from adopting electronic editions. Robinson would remedy these problems with better tools that could make the creation of digital editions easier and their value more apparent to scholars ("Current Issues"). In their explanation of the decision to print the Cambridge edition of Swift in a traditional codex format, Linda Bree and James McLaverty arrive at a far different solution for the same set of problems. Their decision was based on cost, editorial authority, and what uses most readers will make of the text. They point out that "[e]lectronic editions are generally more expensive to produce than print editions" since they must be supported by grants and "the diligent and extensive unpaid labour of academics" (129), and where books can be completed, electronic editions must be indefinitely maintained and updated even when finished in terms of content.[6] While Bree and McLaverty do not question the scholarly utility of the resources available at most electronic archives (different versions of texts, manuscripts, and various social-texts), they do question how many readers will ever utilize these resources or how many will know what to do with them if they do. They envision print editions continuing to be of central use for most readers, with electronic archives supplying different versions, textual witnesses, and other materials to form what Marilyn Deegan and Kathryn Sutherland have called "the hybrid edition" (6).

The intersection of print and digital media is almost certain to continue despite some of the more ambitious claims surrounding electronic scholarly editions. John Unsworth has called sites like the Blake Archive, Rossetti Archive, and many of the other projects that were developed in conjunction with the Institute for Advanced Technology in the Humanities (IATH) at the University of Virginia thematic research collections ("Thematic Research Collections"). The thematic research collection gathers primary digital resources and a "contexual mass" of other materials, and it serves as both a depository and as a site for new research to be conducted (Palmer 357).[7] But as Jerome McGann has written, "personal computers today function most powerfully as scholarly tools when we use them on our desks and in our libraries at home and elsewhere. In those places they get embraced by the more sophisticated, stable, and dispersed network of book technology" ("From Text to Work" par. 19). Stevenson, Johnson, and Fuller all see their editions being used in connection with the Blake Archive. For its part, the Archive often directs users to printed text through its notes, introductions, and bibliographies. Although Nelson Hilton's Blake Digital Text Project provides hypertext editions of the Songs and the Everlasting Gospel, it is noteworthy that both it and the Archive made digitizing Erdman's Complete Poetry and Prose of William Blake an early priority, presumably for the purpose of improving the text through their remediation. As the acknowledged standard text of Blake, Erdman's edition served as the textual foundation of the editions of Stevenson and Ostriker, with Stevenson even sharing the 1971 title page with Erdman. Neither of these editors, however, accepted all of Erdman's decisions. As Stevenson writes in his essay for this volume, he standardized Blake's spelling, punctuation, and initial capitals under the direction of his general editor, F. W. Bateson, who, with Stevenson and Erdman, concluded that Blake's punctuation "is often insoluble" (xii). Ostriker challenged some of Erdman's readings, albeit without specifying which copies or transcriptions she consulted (8). While the Archive is emphatically not based on the Erdman edition (Eaves et al., "Standards" 142), it clearly privileges the text over Bentley, Keynes, and other editions, even though Bentley and Keynes are referred to in the object descriptions alongside Erdman. But whereas these printed editions explicitly adapt and change Erdman's text, the two digitizations of Erdman are silent in how their remediations altered the same text. Figures 1 and 2 provide a screen shot of how both digitizations display the opening of the Songs, and I will assume the reader has access to the print edition for comparison.[8]

Figure 1 Songs of Innocence and of Experience from The Complete Poetry and Prose at the Blake Digital Text Project

Figure 2 Songs of Innocence and of Experience from The Complete Poetry and Prose at The William Blake Archive

The digitization at the Blake Digital Text Project includes line and page numbers in the margins, but does not reference the textual notes, which are only available when using the Project's concordance. More usefully, the Archive digitization provides hyperlinks to the textual notes, which open in a new window. Its concordance is on the same page, which allows viewers to see search terms in their complete context, and it also provides a full index to the work, including the textual notes. In terms of textual reproduction, the stanzas at the Archive have the tendency to collapse into one another, while at the Blake Digital Text Project, the stanza divisions are greatly exaggerated. The Archive digitization editorially adds the plate numbers of the Songs that Erdman chose to leave off but omits Erdman's numbering of the Songs, which followed Blake's 1818 list of the order in which the Songs should appear. Erdman did include plate numbers for the other illuminated books, but in the digitization these are bracketed, giving the misleading impression that the editors of the Archive added them, and in the manuscript works, Erdman's "pages" are often mislabeled "prints." Although neither copy is a new edition, these significant differences mean that the standard edition of Blake now exists in three versions. Do these variations call into question the authority of the letterpress edition itself, and inasmuch as they mirror the variations found in the original prints, do they direct readers to print and digital facsimiles? Or will a reader simply take the digitization at face value, never realizing, for example, how Blake used stanzas? The impact of these digitizations on print culture, moreover, seems minimal since Random House released a "new" print edition in 2008, which added a brief foreword from Harold Bloom, whose dated commentary both digitizations cut without comment.

Such variations are the consequences of remediation, and they underscore how every version of Blake's visual and graphic media into letterpress, printed facsimiles, or digital images will present a different Blake. The inevitable necessity of transforming Blake's works forces us as editors, readers, and scholars to ask, in Morris Eaves's formulation, which Blakes we want and which we do not. In more direct editorial terms, we must ask what in Blake's linguistic, bibliographic, and sociological fields we need and want to represent. As the contributors to this volume all suggest, these questions are inexorably tied to the tools, resources, and perhaps luck available to accomplish these goals.

II

While the following survey of some of the central issues and problems in editing Blake is sketched initially in relationship to the editions of Keynes, Erdman, Bentley, and the Archive, the more reader-friendly editions of Stevenson, Grant and Johnson, and Fuller wrestle with the same issues of how to translate Blake's works into different media and what balance to strike between fidelity and the limitations of both the new medium and the knowledge-base of the reader. In an important article on the history of editing Blake, Eaves identifies the division of Blake's words and images as the critical accomplishment of the first generation of editors that followed Gilchrist. Eaves locates the philosophy behind this division in a pair of statements by William Michael Rossetti and Algernon Charles Swinburne. Rossetti, who had contributed to his brother's edition of Blake and edited his own 1874 volume of Blake's poetry, declared about Jerusalem, "Difficult under any circumstances, it would be a good deal less difficult to read these works in an edition of that kind [i.e., print], with clear print, reasonable division of lines, and the like aids to business-like perusal" (qtd. in Eaves, "Crafting" par. 8). Swinburne, who was one of the first readers to appreciate the words of the prophetic books, dismissed the designs as "mere husk and shell" (qtd. in "Graphicality" 109). As Eaves argues, the decision to move Blake into letterpress was "in response to the (rightly) perceived need for greater legibility" ("Crafting" par. 8). The shedding of this visual husk by Blake's first generation of editors allowed for his "reintroduction into the communication system that he had attempted to resist with his illuminated books" ("Crafting" par. 10). The effect of this simplification was to initiate Blake's slow and steady institutionalization in the twentieth century as a major Romantic poet, but this approach also "systematically fractured Blake's original works" into "disaggregated fragments" that made the works and Blake himself "more compatible with the habits and dominant institutions of modern culture" (par. 4).[9]

But if, as Eaves's history suggests, the fracturing of Blake's works makes him more compatible with "dominant institutions," then the recovery and reunification of his works is, by implication, a subversive act—a chance to redo what Blake failed to do. It may be for this reason that there has always existed a strong countercurrent in Blake editing that seeks to reunite his visual and verbal fragments, and where this has proven impossible, editors and scholars have often pointed to the "wonderful originals" that could not be captured (E 531). In his early typographic transcription of some of the poems, Benjamin Heath Malkin described the Songs as "richly embellished by [Blake's] pencil" (BR 565). Henry Crabb Robinson provided a more detailed description of the Songs and described America and Europe as "works of Blake's poetry and painting" (BR 602). Garth Wilkerson's 1839 version of the Songs verbally described the designs, while C. A. Tulk's small 1843 print run of the Songs took the innovative step of leaving space in its pages for the reader to hand-copy Blake's designs from the original in Tulk's collection (BB 24). All these early excerpts of Blake's text emphasize that what the reader sees is but a fragment of Blake's complete work. The Gilchrist generation initially shared the same tendency, and before Rossetti and Swinburne declared that Blake's works were "Too much" (E 38), editors enthusiastically employed new technologies of reproduction to recover and rescue Blake through remediation. Gilchrist's biography was decorated with inlaid designs copied from Blake's works, while the accompanying edition included electrotypes of the surviving copperplates of the Songs and of the Job engravings. William Muir, John Henry Bellam, and, likely, W. J. Linton all produced high quality facsimiles of Blake's works, some of which passed for Blake's originals (Viscomi, Blake 201-16). While the 1893 William Butler Yeats and Edwin John Ellis edition shared Dante Gabrielle Rossetti's cavalier attitude towards textual accuracy, the edition strove for comprehensiveness by including 296 lithographic reproductions (BB 502). More directly, the Burlington Club brought together many of Blake's visual works in its 1876 exhibition, a means of editing Blake's works that remains indispensible.

The attempt to edit Blake through reproductions was eclipsed by the advent of serious bibliography and reliable print editions in the early twentieth century. Even though this work standardized Blake and divided his words and images, the new bibliographers and editors were acutely aware of the demands of Blake's material forms and the limits of other media in representing them. As noted above, Sampson's 1905 edition was the first to recognize in a practical way that these forms could not be fully represented. With Rossetti and Yeats in mind, Sampson criticized how previous editors had used "Blake's text" as "a sort of poor palimpsest where each new owner has overwritten his own poetry" (vii). Sampson was the first Blake editor to provide bibliographical introductions, alternative readings, and comprehensive textual notes. He was also the first to recognize the "haphazard" punctuation of the illuminated books, which he reluctantly, though explicitly, altered to conform to standardized conventions (viii). Sampson was instrumental in helping Sir Geoffrey Keynes revise his groundbreaking Bibliography of William Blake (1921). The indispensible contribution of Keynes to the editorial and bibliographical work surrounding Blake is best described by Bentley's "Blake's Reputation and Interpreters," but it is necessary to reference some of his accomplishments here for what they say about subsequent efforts to edit Blake. The Bibliography brought together the first comprehensive survey of Blake's corpus, including a verbal description of hundreds of illustrations, reproductions of many other designs, the known references to Blake, and the extant scholarship. Mats Dahlström has pointed to Ross Atkinson's assertion that, like editions, bibliographies "function as simile representations" (30) based on their iconicity.[10] In this sense, Keynes's Bibliography can be seen as one of the most important editions of Blake's work, especially since many of the bibliographic codes it describes have become essential elements of print and digital editions.

Keynes's own letterpress editions were among the first sites where the contest over how to represent Blake's bibliographic codes was waged. The three-volume Writings of William Blake (1925) built on the Bibliography, and it was the only authoritative edition for thirty years (BB 33). Using one copy of the illuminated book as his copy text, Keynes largely respected Blake's idiosyncratic spellings, capitals, and abbreviations, and he included Blake's variants in his textual notes. But in his most infamous decision, Keynes, with the assistance of Max Plowman, standardized the punctuation of the illuminated books "with the admitted risk of sometimes conveying a meaning not intended by Blake" (Keynes, Complete Writings xiii). For the illuminated books, this decision avoided the problem of dealing with the individual copies, while for the works in manuscript, it made them appear far more finished than they were—especially Island and The Four Zoas. Both the Bibliography and the Writings had small print runs, and for the more widely available Poetry and Prose of William Blake (1927), Keynes removed the textual notes and variant readings ostensibly for "the comfort of the majority of Blake's readers" (Keynes, Poetry and Prose ix). It was this simplified version of Blake's text that was the major vehicle of Blake's reception for a generation, including the Beats. Keynes would later write that the removal of the notes and variants from the cheaper edition eliminated much of the "furnace-force of Blake's creative workshop" (Complete Writings x), and he restored and updated the notes for The Complete Writings of William Blake with All the Variant Readings (1957). This edition added line and plate numbers for all the illuminated books based on his and Wolf's Census of the Illuminated Books (1953), which was a move away from what the editors of the Blake Archive describe as "the poem or other work abstracted from its physical medium" ("Editorial Principles"). Although Poetry and Prose contained a number of images, The Complete Writings added images from plates that had backwards writing to the limited number of illustrations found in Poetry and Prose. Keynes explained that such designs were "essential to the understanding of the text" (xiv), and while one sees traces of Rossetti and Swinburne's prejudices in such statements that suggest The Poetry and Prose and The Complete Writings may be the high point for the division of Blake's words and images, Keynes, at the same time, worked tirelessly to bring forward both Blake's words and images in numerous facsimile editions. Most critically, Keynes anticipated the need for facsimile surrogates that could "be placed beside the original book without any fear that the one will suffer seriously by comparison with the other" (414), and his efforts led to the creation of the Blake Trust facsimiles, the most accurate facsimiles available until the advent of digital printing in the 1990s.

In addition to his foresight in creating workable facsimiles, Keynes's bibliographical work facilitated the early computational analysis of Blake's text by drawing together information about different copies of the illuminated books, states of engraving and etchings, textual variants, contextual information, and other collections of data. Erdman's foray into editing Blake began with his desire to add to this computational capacity with a concordance of The Complete Writings (1967). Erdman's Poetry and Prose of William Blake (1965) would ultimately displace Keynes, but until 1982, when Erdman added running line totals to his edition, scholars using Erdman still had to refer to Keynes's Complete Writings as an "interface for locating citations in the Blake Concordance" ("Santa Cruz" 22). Given the subsequent accomplishments of facsimile editing and the individual transcriptions of different copies of the illuminated books and other works available in the Princeton-Blake Trust editions and the Blake Archive, the differences between Erdman and Keynes are less radical than they appeared in the mid-1960s. Erdman's major accomplishment was his effort to represent Blake's own text "as close as possible [. . .] even in punctuation," which involved untangling Blake's "final or preferred readings separated from earlier or deleted or alternative readings or arrangements" (E 709, 1965 ed.). In terms of punctuation, Erdman acknowledged, "In print it is impossible to copy Blake exactly," and he admits to having "wobbled [from his principles] in his transcription" of the prose (E 710, 1965 ed.). But Erdman's approach to Blake's punctuation introduced a central element of Blake's graphic textuality into letterpress editions that previous editors had rejected as frustratingly accidental. By doing so, however, Erdman underscored the inability of letterpress editions to accurately mediate Blake's texts. As he wrote, "In print it is impossible to copy Blake exactly: his colons and shriek-marks grade into each other; he compounds a comma with a question mark; his commas with unmistakable tails thin down to unmistakable periods" (710, 1965 ed.). Erdman based his copy text on the lost copperplates, attempting to recreate them through a collation of the existing prints that could recover Blake's final copperplate intentions. Like Blake's previous editors, Erdman, for the most part, approached the copies of the illuminated books as if they were editions of print books, trying to recreate Blake's ideal authorial intentions and not reflecting the individuality of each book.

Many of the intended and unintended consequences of Erdman's editorial choices and his remediation of Blake's texts were pointed out in the 1984 review of the revised Complete Poetry and Prose (1982) by the Santa Cruz Blake Study Group. While praising the book as "the best available printed edition" (15), the Group highlighted the distortion letterpress editions had on Blake's works: "to change anything that physically appears in Blake's work to an editorial alternative is to 'emend' the text in favor of an editorial line of interpretation" (5). They pointed out instances where Erdman mispunctuated Blake's text in violation of his editorial philosophy, and how his standardization of Blake's line divisions destroyed vital fields of semantic play. Countering W. J. T. Mitchell's idea of Composite Art, which posited, in their view, a false binary between Blake's writings and images, the Group emphasized the interdependently graphic character of Blake's words and images: "There is crucial information of a visual-semiotic nature in Blake's disposition of individual letters, words, sentences and other semantic units on his printed page, and in the visual boundaries that make such disposition possible" (8).[11] A letterpress edition obscures the "graphic potential" (6) of Blake's text, and in ambiguous or polysemous contexts, the letterpress characters themselves force an editorial judgment about authorial intentions. The loss of these features obliterates Blake's status as "the prophet écriture" (4). The Group's review was written at a time when new ideas of textuality and editorial theory were quickly emerging. Both Jerome McGann's A Critique of Modern Textual Criticism and Nelson Hilton's Literal Imagination were published in 1983, and in the following years, Hilton and Thomas Vogler's Unnam'd Forms: Blake and Textuality (1986), Donald Ault's Narrative Unbound (1987), and Molly Rothenberg's Blake's Textuality (1991) all investigated the fluidity and indeterminancy of Blake's texts and the logic of their graphic structures.[12] This scholarship brilliantly revealed dimensions of Blake's work that had been dismissed as accidentals and ephemera but now were shown to be a central component of Blake's method and argument. No edition of Blake has really emerged that fully incorporates this new understanding and appreciation of Blake's textuality into its editorial theory and practice. But, at the same time, the explosion of facsimile editions and catalogues published between the late 1960s and early 1980s suggests that scholars and editors were painfully aware of the limitations of letterpress editions and desired to reflect aspects of Blake's work that textual and editorial theory was just beginning to enunciate.

Viscomi has faulted these print facsimiles for "point[ing] to the original rather than reproduc[ing] it" (Kraus 148). But these editions are instructive because they show editors struggling to theorize and to represent Blake's complex bibliographic codes without fully understanding his production methods and the formal limitations of the codex. Erdman's Illuminated Blake (1974) and David Bindman's Complete Graphic Works of William Blake (1978), for example, offered an illusory completeness fostered by the limitations of print. Yet their black and white images of relatively low quality seemed impossibly rich in visual and bibliographic detail in comparison with letterpress editions. As is now obvious, the codex form forced fatally misleading representations of Blake's work. Martin Butlin's still indispensible Paintings and Drawings of William (1981) shows that the same is true of Blake's productions as an artist since the volume closely crops Blake's designs. With the quality of their reproductions, all three of these books must convey essential bibliographical data through description rather than the representation of the artifact. Given these limitations, it may not be coincidental that this era saw editors experimenting with remediating Blake's work through other media, such as slides, microfilm, and filmstrips. Interestingly, among the most accomplished parts of this print archive were facsimile editions of Blake's manuscripts, which may have achieved a standard that excelled the reproductions of the illuminated books because more theoretical and practical models existed. Erdman's photographic and typographic facsimile edition of Blake's notebook (1973) was a magnificent achievement of diplomatic editing that provided black and white photographs of each leaf with transcriptions that followed the layout, font size, emendations, and deletions with a level of accuracy that will be very difficult, if not impossible, to replicate electronically. In 1963, G. E. Bentley, Jr. published a facsimile edition of Vala or The Four Zoas that provided full-sized black and white images, detailed transcriptions, and a wealth of bibliographical data that greatly influenced later conceptions of the poem's status as a manuscript work. Bentley's facsimile has been supplemented by Erdman and Cettina Tramontano Magno's facsimile edition and commentary on The Four Zoas (1987), which used infrared photography to recover wild and often erotic scenes that Bentley's reproductions were not able to capture. That same year, Michael Phillips published his facsimile edition of An Island in the Moon, which provided a detailed diplomatic transcription of the work that challenged previous readings of the work in letterpress editions and made apparent the textual thicket from which some of The Songs of Innocence emerged.[13]

Bentley's other contributions to the print archive were just as significant as Vala. Beginning with his and Martin Nurmi's Blake Bibliography (1964) and culminating in Blake Books (1977, with updates), Bentley revised and expanded Keynes's Bibliography, which had become within a few years of its publication a rare book that "very few [could] hope to possess" (Binyon v). Blake Books remains as indispensible as ever. Conceived as an edition of Blake, the book represents Blake's materiality and his relationship to print culture in comprehensive ways that no electronic archive could ever hope to match, and its bibliography of scholarship is so complete that it is almost a twentieth-century reception history. Because it anticipates the goals of social-text editing laid out by McKenzie and McGann, Blake Records (1968, with updates) stands as an innovative edition of Blake that traced his presence in a legion of social networks. Together, these works greatly expanded the editorial and interpretative fields associated with Blake, and in many ways they have constituted these fields so pervasively that scholars who have railed against the dominance of Erdman have scarcely perceived Bentley's remediation of this material.

Bentley's insistence on Blake's materiality and visual dimensions is apparent throughout Blake's Writings (1978), which was a decisive step in the fusion of letterpress and facsimile editions of Blake. His purpose was "to present Blake's writings in a form as close to his originals as type will permit," and he included "all the printed designs which are of considerable size and significance" (xxxviii). These reproductions were of far higher quality than those found in Erdman's Illuminated Blake, though they were similarly reduced and standardized in size. Bentley was not able to reproduce every page of every poem, and he typically omitted those plates Vincent De Luca has described as a "wall of words."[14] Despite the hundreds of designs Bentley reproduced, however, his text was not a transcription of these plates, and like Erdman, he envisioned the lost copperplates as his copy text. Rather than collating copies of existing prints, Bentley favored "copies in which the etched text is clear, uncoloured, and not clarified by hand" (xlii). Typically, these copies were posthumously printed and reflected the final state of the copperplate "unaffected by the author's colouring or changes of mind in the process of printing or correcting" (xlii). Nonetheless, Bentley joined with Erdman in isolating and idealizing Blake's work on the copperplate from the other stages of production. In punctuation, Bentley tried to strike a balance between Keynes and Plowman's silent editorial intervention and Erdman's attempt to record every mark by showing where he added, subtracted, or emended Blake's punctuation through typographical emphasis. He provided "essential" marks in half brackets; omitted "redundant full stops," marking the omission by italicizing the final letter of the preceding word (his "discontent" for Blake's "discontent."); supplied the correct punctuation in place of the full stop, again marking the change with italics ("scorn?" for "scorn."); and capitalized names and beginnings of paragraphs that Blake wrote in lower case (xliv). Bentley's representation of Blake's text and punctuation elicited a thoughtful commentary by E. B. Murray who used the difficulty of Blake's punctuation to question the applicability of the copy-text paradigm to Blake. Pointing out the substantial disagreement between Erdman (1965) and Bentley in marking the unique copy of The Song of Los, Murray proclaimed that "Blake's pointing is often so ambiguous in its appearance that there is no defining it except arbitrarily" (149). Instrumentally, Murray challenged the idea that an editor could choose an ideal copy-text or produce a consensus text that would best reflect Blake's intentions, and he called on scholars to pay greater attention to the production of the illuminated books.

III

The call to investigate Blake's production methods would be best met by Morris Eaves, Robert N. Essick, and Joseph Viscomi, who together tempered the unsubstantiated claims for Blake's textuality by focusing on his production methods in their historical context. Eaves, who has always shown an interest in Blake's use of technology, positioned Blake within the overlapping discourses of art, British nationalism, and technologies of reproduction in The Counter Arts Conspiracy: Art and Industry in the Age of Blake (1992). In William Blake: Printmaker (1980), Essick firmly situated Blake's biography and illuminated books in the tradition of professional engraving, and in a subsequent series of articles, he challenged some of the untenable conclusions of textual critics who removed Blake's works from the site of their material and historical production and privileged Blake's aesthetic theories over his practice.[15] It was Joseph Viscomi's explanation of Blake's production methods in Blake and the Idea of the Book (1993), however, that had the most direct impact on editing Blake. Viscomi shows that Blake did not produce the books for individual customers one at a time but printed multiple copies of the books during the same printing session. Viscomi terms these copies an "edition" (183) because they reflect common production procedures. Each edition, moreover, had an "ideal copy," a term Viscomi borrows from editorial theory where it refers to a book with all its leaves intact that would be sold by a publisher (179). For Viscomi, an ideal copy is how Blake envisioned the editions at the moment of production, which included those features shared by all the copies of the edition. Ideal copies "can be used to determine Blake's intentions for printing sessions, at least in regard to the number, if not also the order, of plates that a book was to have when produced" (179). A new edition, or print run, could represent a new reinvention of the book, but not each copy of an edition since true "versions," as Viscomi employs the term, resulted from "different periods of production" (178). Viscomi applies his knowledge of Blake's production methods to defend Erdman's and Bentley's practice of privileging the copperplate text on the grounds that it was during the preparation of the copperplate that Blake showed his last real interest in "verbal structures" (182).[16] Viscomi's dating of the illuminated books enables editors to know Blake's intentions at a given moment, and he argues that the changes in color and composition among the various copies of one edition do not substantially change a poem to the same degree that a new system of production does (182). Viscomi's advice for the transcription of Blake's problematic punctuation was to record it "as it reads to the eye," with variants being recorded according to the edition and arranged chronologically, but he gives no clear suggestion on how to transcribe the marks that have no counterpart in type.

These investigations into Blake's production methods coincided with further developments in textual studies and the advent of new media. These changes radically altered notions of textuality and the assumed primacy of the book as the vehicle of scholarly editions.[17] Blake scholars who had long been frustrated with the remediation of Blake in print now had material alternatives. Nelson Hilton at the Blake Digital Text Project began to develop the digitized Erdman and the hypertext editions of the Songs and Everlasting Gospel with the belief that the new platform would reveal much about the nature of Blake's textuality.[18] Eaves, Essick, and Viscomi had their first opportunity to put their editorial theories into practice in two of the six Princeton-Blake Trust facsimile volumes. While this series was a milestone in terms of the quality of its images, the transcription of each individual plate, and its scholarly introductions, annotations, and other notes, the dissatisfaction of Eaves, Essick, and Viscomi with the platform of print for representing Blake's works led them to IATH and their contact with its director John Unsworth. With the initial assistance of IATH, Eaves, Essick, and Viscomi gave birth to new ways of remediating and editing Blake. The Archive's basic idea of providing vibrant color images of different copies that were remotely accessible radically transformed what it meant to read, study, and edit Blake.[19] The success of the Archive in meeting many of the mundane challenges laid out by Robinson, Bree, and McLaverty is due to the tireless work of its editors and project managers in obtaining grants, working with technical staff, training and directing project assistants, and dealing with the scores of different institutions and individuals holding Blake's works, even as they wrestle with the more traditional problems faced by editors in what was and remains a new and evolving digital environment. The individual transcription of each object allows users to focus more than ever before on the unique copy before them. Each of these objects can be viewed in several different ways. 100 dpi jpeg reading images can also be viewed at their actual size using Inote, while 300 dpi jpeg images can "provide more detail than even unmagnified originals" (Viscomi, "Digital" 32). In other words, digital media allow a host of bibliographical data to be transmitted by the image itself. The current comparison feature of the Archive allows for different versions of the same plate to be juxtaposed, and when the Archive finally launches its Lightbox software, viewers will be able to compare any set of visual material they wish. The Archive also offers many computational options. Its transcriptions are fully searchable, and the editors have developed a verbal tag system for searching Blake's images. In addition to its electronic version of Erdman, the Archive also provides comprehensive textual notes describing the work's material history; introductory essays on Blake's life, mythology, and engravings; bibliographies; contact information for the holders of Blake's works; and a detailed account of its own editorial methodology, history, and reception. In addition to the illuminated books, the Archive includes many of Blake's drawings, paintings, and engravings. As the editors write,

By incorporating as much of Blake's pictorial and literary canon as possible—with both images and texts organized, interlinked, and searchable in ways that only hypermedia systems will allow—the Archive would for the first time give scholars and students access to the major intersections between the illuminated books and Blake's other creative and commercial works. That is to say, by exploiting new information technology to deliver the historical, technical, and aesthetic contexts necessary to study Blake as printmaker, painter, and poet, the Archive would encourage a deeper, more responsible understanding of his aims and methods, which have been regularly misunderstood and misrepresented. ("The Plan of the Archive")

The wealth of materials now available at the Archive makes it possible to fuse Blake's different artistic identities and modes of production, and, as Lee and McGhee and Van Kleeck detail in their essays, the Archive recognizes the need to adopt new editorial philosophies and practices to represent these different works.

The practical accomplishment of the Archive as a useful "edition" of Blake's work has been accompanied by a running set of meta-reflections by its editors, project managers, and project assistants on the novel experience of editing in digital media. Many of the editorial choices of the Archive have emerged from a combination of long email chains that wrestled with theoretical concerns and the dictates of technical capabilities that speak to Martha Nell Smith's claim that "editing is a kind of encoding and coding is a kind of editing" (314). As Lee and McGhee show, many of the technical issues confronted by the editors and staff of the Archive are a fundamental part of the editorial process when working in digital media. Eaves has described how the idea of an electronic edition initially held the promise of an Osiris-like reunification of Blake's words and images, but the experience of editing the Archive has taught him that the site "is at least as much about recapitulation and recycling as about restoration—and as much about fragmentation as integration" ("Crafting" pars. 4 and 5). Because digital editing "conforms to the curve of its technologies" even more than print (par. 28), it confronts editors and projects with a "constant threat of change and obsolescence" that is ultimately "unnerving" (par. 29). Despite the innumerable differences in editing for print and electronic platforms, Eaves also stresses the continuity the Archive has with the tradition of editing Blake in print:

Instead of the fresh, untrammeled view of the original Blake that one might imagine, the work presented in the Blake Archive is enmeshed in a framework of supplementary information and optional views defined over the last 200 or so years by Blake's most talented and resourceful sponsors. The superconsolidated array and the reorientation to visual reproductions of Blake's original documents are new, at least on this scale (if ungainly), and the scholarly opportunities they offer altogether are unprecedented. But the elements are, at bottom, inherited. The entire editorial grid, reproductions and options alike, is further restricted by interlocking technical compromises imposed by the present, rather severe, limits of memory, bandwidth, software and hardware design, institutional requirements, and our own editorial imaginations. (par. 25)

The debt of the Archive to the tradition of print underscores McGann's point that computers are most effective when placed in a network of books. Moreover, as Johnson's essay in particular will make clear, as does her experience with the Blake Videodisc Project,[20] the constraints and difficulties articulated by Eaves are embedded in every editorial venture regardless of its media.

Yet the difficulty of obtaining funding and of crafting agreements with holding institutions makes it hard to imagine another site emerging that could challenge the Archive in the same way that letterpress editions have challenged one another. For better or worse, successful thematic research collections like the Archive are in unique positions to shape their authors in ways that far exceed the authority of the most popular print editions, and the exceptional influence of the Archive demands a detailed understanding of how it remediates Blake. The most obvious remediation of the Archive is its vast array of digital reproductions. Kathryn Sutherland has challenged more grandiose claims for digital surrogates made by scholars like Martha Nell Smith. Quoting Matthew Kirschenbaum,[21] Smith argues,

The ability to provide artifacts for direct examination (rather than relying on scholarly hearsay) has altered the reception of humanities computing in the disciplines of the humanities so that skepticism is "at least replaced with more to-the-point questions about image acquisition and editorial fidelity, not to mention scholarly and pedagogical potential." (Smith, "Electronic" 310)

Troubled by the easy equation of high quality digital images with "artifacts" that circumvent scholarly mediation, Sutherland fears that most users will not question the fidelity of the images they see, believing the commercial and scholarly rhetoric that suggests "the computer is a totally mimetic space unshaped by the constraints of its own medium" ("Being Critical" 17-18).[22] Sutherland's fears are manifest in Kenneth Price's recent declaration that "The electronic edition can provide exact facsimiles" ("Electronic Scholarly Editions"). As Sutherland insists, the "electronic instantiation" of these facsimiles is severely under-theorized ("Being" 18). For their part, the editors of the Archive have been very explicit about how their images are obtained and what they can provide the user:

[O]ur images are not intended to be "archival" in the sense sometimes intended—virtual copies that might stand in for destroyed originals. We recognize that, if we are going to contribute to the preservation of fragile originals that are easily damaged by handling, we must supply reproductions that are reliable enough for scholars to depend upon in their research. Hence our benchmarks produce images accurate enough to be studied at a level heretofore impossible without access to the originals. (Eaves et al. 138)

While this benchmark has been unquestionably met, anyone who has seen these originals know that the Archive images flatten, obscure, or simply remove such features as vibrant color; gold plating; fine ink work; the wide margins of a page on which the print may be strangely cockeyed; bindings; stab holes; new mountings; museum stamps; editorial and curatorial notes; material on the back of the page, print, or painting; and a host of material contexts in which engravings, illustrations, paintings, or annotations appeared. Again, the editors of the Archive have not shied away from these issues but explored them in great detail. Discussing how the Archive crops the sheet surrounding Blake's designs, Joseph Viscomi has acknowledged that the process "discards bibliographical information, [and] thus fails to represent the artifact itself" ("Digital" 34). The reasons for doing so, he writes, are "more technical and pedagogical than economic and aesthetic":

Recording the image on the film as large as possible reduces the scaling ratio between reproductive source and original, which ensures greater accuracy in the digital image [. . .]; cropping out the image's margins significantly reduces file size, and, for most images, enables the image to be shown within the Archive's interface on monitors 17" or larger without scrolling. ("Digital" 34, 35)[23]

The decision to crop the image of the artifact stemmed from an imminently practical set of concerns related to technical resources and the desire of most users to have an image of Blake's design that is as large as possible. But the practice of cropping moves the Archive farthest from its emphasis on "the physical object" ("Editorial Principles"), and it is probably not coincidental that here the Archive closely follows the practices of print facsimiles, which share a similar set of technical and pedagogical concerns. As technical capabilities inevitably change, users will be haunted by decisions demanded by the moment. Eaves has recently described some of the current technical difficulties facing the Archive in achieving its editorial vision, focusing particularly on the problem of using jpeg images, which cannot capture the clear, crisp engraved line, and the long struggle by the Archive and others to provide a way of searching images without recourse to a verbal-tag system. These technical problems show the real and frustrating limits of working with a host of tools that are not completely one's own. While the image of the design can provide more detail than the original, the remediation of the plate will invariably shape the kind of research that can be performed. The fact that many museums, libraries, and private collectors are restricting access to Blake's works also creates a situation in which the digital surrogates may have to fill in for many of the functions of the originals not foreseen by the editors, a need that may demand access to or the creation of images that contain more information about the complete artifact and its material context. Solving these problems will also invariably involve the holding institutions, which remain necessary partners when their collections are being used. As both Eaves and Nelson Hilton have stressed, the accessibility of Blake's images in digital media has not meant their free circulation and use since both editors and scholars must still deal with issues of copyright and accessibility, with Eaves stressing how the Archive is contingent on "formal and informal" agreements with these institutions ("Golgonooza" par. 9, "Picture" par. 5).

Many other implications of the Archive's remediations are explored in the essays by Lee and McGhee, Van Kleeck, and Ripley in Editing and Reading Blake. In "'The productions of time': Visions of Blake in the Digital Age," Lee and McGhee trace the editorial and technical evolution of the Archive up through its current work on the manuscript satire, An Island in the Moon (1783-85). Recounting their work marking the Island in the Moon manuscript, Lee and McGhee make the crucial point that not every physical marker, linguistic code, or textual variation can be realistically marked, leaving elements that some users will consider vital out of the representational and computational scope of the Archive.[24] Lee and McGhee also show how the recent transition to XML impacts editing practices of the Archive and how the release of the fifth version of the TEI Guidelines allowed the editors and project assistants to encode the alterations Blake made to the Island manuscript with far greater nuance. Their essay shows how even though electronic editions are haunted by editorial choices dictated by technical and other practical considerations, the perpetually changing nature of electronic tools will enable the Archive to reflect different aspects of Blake's works and grow as a project in relationship to the needs of the scholarly community that uses the site. To address the unique problems presented by manuscripts, the Archive developed a new system of transcription, and Lee and McGhee provide a valuable description of this new color-coded system. Their account is a clear demonstration of how editing in digital media is contingent upon a host of tools whose capabilities and limitations are not under the control of the editor and how editors must improvise to make these tools work.

In "Editioning William Blake's VALA/The Four Zoas," Van Kleeck examines the editorial history of a manuscript far more complicated than Island, Blake's VALA or The Four Zoas (1797-1808?). Beyond its two titles, the work's legendary host of editorial difficulties includes the palimpsest of revisions found on several pages; the uncertain arrangement of text, pages, and nights (including its famous two night sevens and the unnumbered "Night the [Second]"); the various scripts in which Blake wrote different sections of the text; the manuscript's appropriation of the Night Thoughts proofs, a proof of Edward and Elenor, and other scrap pages; and more than one hundred designs. Building on Donald Reiman's idea of versioning,[25] Van Kleeck coins the term "editioning" to describe how each edition presents a different view of a work, and he surveys the editorial representations of the manuscript/poem by Keynes, Margoliouth, Erdman, Bentley, and Stevenson in masterful detail, providing a valuable and lasting contribution to the editorial history of the work. Van Kleeck also considers the interpretative problems arising from having to deal with print editions and facsimiles in the absence of easy access to The Four Zoas manuscript, which is held tightly in the British Library vault. Van Kleeck describes his own experience with the artifact and how this experience affected the electronic edition of the poem that he created for his dissertation, which will serve as the foundation for the Archive's edition.

In "The Delineation Editing of Co-Texts: William Blake's Illustrations," Ripley suggests that Blake's illustrations of other authors offer an ideal site for exploring D. F. McKenzie's and Jerome McGann's notions of social-text editing. Ripley argues that when the Archive edits out the literal presence of Blake's source texts in its transcriptions of Blake's designs or scatters the contents of what was once a bound, printed volume across its different material categories, the Archive crops Blake's works in very significant ways that replicate notions of Blake as an autotelic artist. While these choices are the result of practical editorial decisions, the decisions still distort and destroy the original coherence of Blake's collaborative works and their existence as social texts. Ripley suggests that the texts and illustrations are necessary co-texts that must be available in the same editorial field for the illustrations to be read correctly. Building on the recent work of Joseph Viscomi on Blake's virtual designs and of Saree Makdisi on Blake's graphemes, Ripley offers a social-text model of editing Blake within his material and social contexts that uses Blake's theory of the line and which would look to models such as E. L. Ayers and W. G. Thomas's The Valley of the Shadow: Two Communities in the American Civil War and other IATH projects.

IV

Future print and digital editions of Blake will have to recognize needs and capabilities not met by the Archive. For all its current computational power, the Archive still lacks many of the editorial tools called for by Robinson. Susan Hockey, who emphasizes that "a digital image is a surrogate, not the real thing" (371), has argued for organizing digital scholarly editions in new ways not dictated by the model of the printed page. As large as Blake Books and its updates are, a digital platform could allow for a host of additional bibliographical and contextual information that would facilitate computational analysis of many different features and topics. Eighteenth-Century Collections Online and Google Books have shown the potential of this direction because they allow for full-text searches and provide a much broader range of paratextual and contextual information than a print bibliography, even if the remediations of this material can be problematic. Such tools could be developed independently or in conjunction with the Archive.[26]

Another direction can be found in Edward Vanhoutte's call for a minimal edition, which might be termed the contrary to the thematic research project:

The minimal edition is a cultural product that is produced by the scholarly editor acting as a curator or guardian of the text, whereas the maximal edition is an academic product in which the scholarly editor demonstrates his/her scholarly accuracy and scrutiny. (100)

Vanhoutte sees the reading edition as the desideratum of electronic editions. Scholars like McGann would rightly counter that the end of textual criticism is not accuracy but precise, historical interpretation that electronic archives have helped to enable ("Monks" 66). But the wider point about the need for a minimal edition as a cultural product remains relevant, especially when one reflects on the impact of Keynes and Erdman. Through its stress of each copy of the illuminated book as well as its digitization of Erdman, the Archive has hitherto largely avoided the problem of having to produce a minimal edition of its own and to wrestle with the theoretical and practical implications that the edition would create. But as Lee and McGhee relate, Blake's manuscripts resurrect this problem for the Archive in several ways. The diplomatic transcriptions of individual copies of illuminated books have the added value of also being reading versions that are more accurate than the collated versions of letterpress editions. Applying the same diplomatic editorial philosophy to manuscript pages will reflect its textual contradictions and incoherence. In ways that recall Erdman's description of Blake's "shriek-marks" (E 787), Lee and McGhee describe how the last page of Island resists the logic of diplomatic transcription because of the indeterminate nature of Blake's drawings and letters. At some point, as they suggest, the editor needs to decide whether a mark is a character capable of transcription or an image to be described. Van Kleeck and Stevenson both emphasize how editors of The Four Zoas have long recognized that the poem requires an editorial intervention, even with Ault's magnificent reading of the poem as a manuscript largely following the structure given by Erdman's 1982 arrangement. The Vision of the Last Judgment, Public Address, and many of the notebook poems have been recovered through editorial interventions, while, as shown recently by Jason Snart, Blake's marginalia has a complex intertextual and graphic relationship with the books in which they were written.[27] In short, the absence of a minimal edition means that the Archive will not be presenting legible texts for some of Blake's most famous works.

The Archive also lacks space for editorial interpretation and contextual annotation. With the exception of the hypertext editions of the Songs and the Everlasting Gospel at the Blake Digital Text Project, which have not been maintained and contain many dead links, interpretative editions remain the province of print facsimile and letterpress editions, particularly those created by the contributors to this volume, Ostriker, and the Blake Trust Series. In addition to his work as an editor of the Archive, Essick has continued his work in producing print facsimiles, issuing editions of the Virgil engravings (1999) and the Huntington's copies of the Visions of the Daughters of Albion (2002), the illustrations to Paradise Lost (2004), and the Songs of Innocence and of Experience (2008). The copies of these illuminated books (E in both cases) are not currently available at the Archive, and Essick reproduces features of their bibliographic codes that cannot be imitated by the electronic Archive. The facsimile of Visions, for example, is printed on both sides of the leaf to imitate Blake's original printing (Essick 15), an innovation that had been urged by Viscomi (Book 180). Such choices reflect the truth of Sutherland's recent insistence that there remain many aspects of the book that cannot be represented electronically ("Being Critical" 21), and a printed book's mediation of Blake's work may be closer in some aspects than a digital image.[28] Reproducing an individual copy allows Essick to perform a specific reading of the copy and "not a generalized interpretation of all copies or of some hypothetical 'standard' copy" (26). Exhibitions and catalogues are even more dramatic examples of Blake's originals with useful contextual information as seen, for example, in Robin Hamlyn and Michael Phillips's beautiful catalogue for the 2000-01 Tate exhibit or Martin Myrone's recent catalogue of Blake's 1809-10 exhibition, Seen in My Visions.[29]

In light of the ability to produce such reliable print and digital facsimiles, the letterpress edition may seem too minimal. But part of the continuing success of the annotated editions of Stevenson, Grant and Johnson, and Fuller without a doubt stems from their ability to present useful information to their readers through the efficient exploitation of print and its scholarly apparatuses. Claire Warwick has praised the technology of print in ways reminiscent of Rossetti: "Reading a printed text is clearly a subtle and complex analysis technique. It is therefore not surprising that scholars have made the assumption that digital resources and computational techniques that simply replicate the activity of reading are a pale imitation of an already successful technique" (374). Stevenson, Johnson, and Fuller skillfully utilize the conventions and technologies of print, doing their readers a great service by mediating both the knowledge accumulated in an unwieldy critical tradition that has made sense of Blake and the complex bibliographical and typographical issues for which most general readers do not have the patience or training. These remediations reflect editorial theories and practices just as complex and rigorous as those of Keynes, Erdman, Bentley, or the Archive as they wrestle with both the demands of Blake's works and the requirements of different audiences.

In "The Ends of Editing," Stevenson concentrates on the practical outcome of editing, which is to provide the reader with "an accurate and useful text." In the three editions of The Poems of William Blake (1971), retitled Blake: The Complete Poems for the 1989 and 2007 editions, Stevenson wrestled with these sometimes conflicting dictates in light of what readers need and what the text demands. As he observed in the 1971 preface, no annotated edition of Blake's poetry had been produced since Sloss and Wallis's 1926 edition of the prophetic books (xi), and Stevenson's choice to do so was influential, since, by the end of the decade, two additional annotated editions (Ostriker and Grant and Johnson) would be published. Where the Bloom commentary was an "explicatory discussion rather than [a] detailed annotation," Stevenson aimed to "annotate all of Blake's poems line by line and detail by detail," providing "essential details of fact and background" (xi). In the 1971 edition, Stevenson supplemented the annotations with a general introduction, headnotes, maps of the Holy Land and London, and five black and whites plates. In textual matters, as noted above, he worked closely with both Erdman and Bateson in producing the 1971 edition. As Stevenson writes, Bateson insisted that no grammatical or typographical difference should interfere with the "sympathy" between Blake and his readers. Even if Bateson's editorial principles dictated the wince-inducing spelling of "Tiger," the standardized spellings and punctuation greatly aided readers. It was this service that distinguished his editions from those of Erdman and Bentley who struggled to "see what words and points Blake actually put down on paper." In the two subsequent editions, Stevenson allows more of Blake's idiosyncrasies to stand, but he still justifies the standardizations of his text on the grounds that his edition is "designed to be widely, and fluently, read" (xiv, 2007 ed.).

The serious revisions Stevenson has made in each successive edition underscore the relevancy and vitality of his work for both general and specialized readers. He opens the 1989 edition with the observation that "much has changed in the Blake world, as well as in the world outside" (xi), while in 2007, he describes how he "scoured and revised" the headnotes and footnotes in light of Blake scholarship and editorial theory over a period of nearly forty years (xiv). While continuously updating his edition, Stevenson also stresses in his essay that he constructed his notes around simple facts (e.g., place names of London), facts that have wider significance for Blake (e.g., Beulah), and those images intertwined with other images (e.g., the "Three Classes" of Milton). Stevenson's standardizations and annotations do not mean that his edition eschews textual difficulties. As Van Kleeck describes in detail, his 1989 arrangement of The Four Zoas revealed a tenable alternative vision of the poem, and in the 2007 edition, he made such major changes as adopting the plate order of the later copies of Milton and including the tractates because they added to the reader's understanding of Blake. With the addition of color plates for the 2007 edition, Stevenson also included more of Blake's designs, with the goal of illustrating how Blake "repeatedly used certain [visual] motifs" (xiii). In line with his standardization of Blake's text, his images are closely cropped, and, with the exception of Urizen plate 5, they reveal nothing of the wider page upon which they appear. But Stevenson remedies this remediation by referring readers to the Princeton-Blake Trust Series and the Blake Archive. Such decisions are well considered in light of new scholarship and reflection regarding the needs of the reader, but, as Stevenson notes in his essay, Blake often forces an editor to make decisions that cannot ultimately be justified by the text itself. In the end, Stevenson admits that "There is no simple formula for" editing.

Fuller's William Blake: Selected Poetry (2000, 2008) also standardizes spelling and mechanics of his selections of poetry, supplying the reader with a general introduction to Blake, useful headnotes and annotations, a bibliography, and nine illustrations. To supplement these illustrations, Fuller, in a unique move in modern editions, verbally describes the most important designs, as well as pointing readers to the Archive. In "Modernizing Blake's Text: Syntax, Rhythm, Rhetoric," Fuller rejects the general push in Blake editing toward replicating the artifact and embraces the idea that the primary mission of an editor is to "modernize" Blake's text for the audience. Fuller stresses that all editing is a form of construction, and he argues that a modernizing editor must assume the role of the ideal reader. Confronting Essick and Viscomi's description of Blake's conflicting practices regarding the state of his text, Fuller points out that editors must decide when Blake's idiosyncratic punctuation and other bibliographic figures present great meaning and when they are the result of Blake's production methods or cavalier attitude towards some particulars.[30] As an example of the former, Fuller calls attention to how Blake's punctuation creates "structurations" in Blake's verse, such as "accreted parallelisms," which are hidden by diplomatic transcriptions. While his edition does not do so, Fuller points out that the modernizing editor could reveal such structures to the reader in the layout of the poem, calling attention to the wider patterns of the work in ways reminiscent of McGann's discussion of editorial defamiliarization. Fuller is also more attuned than other editors of Blake to the rhythm of his language, making a persuasive argument for marking the accented -eds and retaining the elisions as -'d or -d to aid the reader in replicating Blake's rhythm and sound and following a practice laid out by Blake's first professional editor, John Sampson.

In terms of book production alone, both editions of Grant's and Johnson's Blake's Poetry and Designs (1979, 2008) have been beautiful fusions of Blake's images and words that give new readers a useful introduction to Blake's visual and graphic productions while also offering lucid headnotes, textual notes, annotations, maps, critical essays, and bibliographies. In "Contingencies, Exigencies, and Editorial Praxis: The Case of the 2008 Norton Blake," Johnson chronicles the difficulties of revising the work in light of developments in Blake scholarship, editorial theory, and new methods of editing. In the 1979 edition, she and Grant constructed their text from copies that were available from facsimiles, so that readers could compare the transcription with the reproduction, and unlike other editions, they tried to follow Blake's line spacings and indentations, going so far as to replicate the ornamental stanza divisions in their text (xliv). As Johnson writes, she and Grant, in preparing the new edition, confronted a range of editorial questions that they had simply not asked before, and they employed the new resources and information provided by the Archive, The Blake Digital Text Project, the Princeton-Blake Trust series, and Viscomi's Blake and the Idea of the Book to grapple with Blake's texts anew. The base text for the 2008 edition originated with the "earliest edition that contains all or most plates appertaining to a given work," which would be compared to the copies at the Archive, Erdman, Bentley, and the Blake Trust series (599). These comparisons did not alleviate the central difficulty of translating Blake's marks into letterpress (600), and while their edition does not standardize Blake's punctuation to the degree that Stevenson and Fuller do, it still must struggle to "balance accuracy with readability" (602). In practice, this has meant removing "punctuation that seriously interrupt units of thought" and deciding not to replicate Blake's line breaks, even when they offer important arguments of their own (as in the famous example of "book of me / -tals from The Book of Urizen) (601).

If these decisions standardize Blake too much for some readers, the 2008 edition stresses that it is to be used "in tandem with the magnificent William Blake Archive" (xi). The edition provides a far more elaborate and accurate system of reference to the plate and object numbers used by Keynes, Erdman, Bentley, and the Blake Archive, which facilitates and encourages cross-references by the reader to other editorial vision of Blake. In collating the text of different copies, the 2008 edition presents lines from other copies but strikes them out in a manner that retains their legibility but clearly indicates that they were missing from the copy at hand. As the title of the edition underscores, Blake's visual work remains central, with many of the images positioned within the text itself, remediating a new and fruitful interplay of the words and designs. The new edition adds to this effect by more skillfully allowing the white space of the page to bleed into the designs. While the color plates are more vibrant, it is worth noting that the newly remediated images consistently omit Blake's page numbers, even though they were present in the "lower quality" image in the 1979 edition. Such problems underscore how even ostensible improvements are editorial changes with immense consequences. With the new edition containing sixteen color plates, 86 black and white designs integrated in the text, and the entirety of Jerusalem, Johnson illustrates how real world exigencies enter into editorial decisions that were worked out in terms of finances and her interactions with the publishers. As troublesome as these exigencies can be, Johnson stresses that they also create unforeseen opportunities, as seen in her account of why and how the new edition has colored maps on its end papers.

Featured with Johnson's essay is an appendix listing the errata of the 2008 edition, a special benefit to readers of Blake, Praxis, and the Norton volume. In this list, Johnson has continued the melding of the print and digital resources that makes Blake's Poetry and Designs such an important book in envisioning the future intersection of these platforms. New print and digital editions of Blake will have much to learn from the work of Stevenson, Johnson, and Fuller, whose integration of the general reader into their editorial vision should be the model for any future uses of immersive textuality in approaching Blake. But, for the present moment, these current editions of Blake demonstrate that print editions can engage with the resources offered by the digital age and remain highly relevant to general readers and scholars alike.